Free tool, no account needed
See which AI bots can read your site, and what that means for you. We check your robots.txt, your page and how your server answers each bot.
AI companies send different bots for different jobs. Blocking one is not the same as blocking another.
They build the index AI search answers come from: OAI-SearchBot for ChatGPT search, Claude-SearchBot, PerplexityBot, and Googlebot and bingbot for Google's and Microsoft's AI answers. Block one, and that product can't show or link your pages.
They collect pages that may be used to train AI models: GPTBot, ClaudeBot, meta-externalagent, Amazonbot and CCBot. Blocking them opts you out of future training. It doesn't remove you from AI search.
They open a page the moment someone asks about it: ChatGPT-User, Claude-User, Perplexity-User. OpenAI, Perplexity, Meta, Amazon and Google say some of these may ignore robots.txt, because a person asked. Anthropic says all its bots follow robots.txt.
These are not crawlers. They are names in robots.txt that tell Google and Apple whether pages they already crawl may train their models. Google says Google-Extended does not affect Google Search, so it does not change AI Overviews or AI Mode.
So blocking a learner keeps your pages out of future training, while blocking a searcher keeps them out of that AI product's answers.
robots.txt is a sign on the door, not a lock. It asks bots to stay out, but it can't stop them. The lock is your server, host or CDN: a firewall rule, a security plugin or a bot setting can turn bots away while robots.txt welcomes them.
Cloudflare is a common example. Since 1 July 2026, every Cloudflare plan can block AI search, agent and training bots separately. Since 15 September 2026, new Cloudflare sites block training and agent bots on pages with ads by default.
So if our test finds a bot turned away, check your firewall, bot protection and CDN settings first.
Cloudflare's bot changelogWhat we can't see: real bots come from their company's own IP addresses. A firewall that checks those can treat them differently from us, in either direction. That's why the server test is a signal, not a guarantee.
A check sends about 16 small requests, at most 3 at a time, and only when you press the button.
The robots.txt standard (RFC 9309)No. GPTBot collects pages that may train OpenAI's models. ChatGPT search uses OAI-SearchBot, and OpenAI says the two settings are independent.
No. Google says Google-Extended does not affect Google Search. AI Overviews and AI Mode are part of Search, so they follow your Googlebot rules. To limit what appears there, Google points to nosnippet, data-nosnippet, max-snippet and noindex.
Most bots that collect or index pages say they do. Several bots that open a page because a user asked say robots.txt may not apply: ChatGPT-User, Perplexity-User, meta-externalfetcher, Amzn-User and Google's user-triggered fetchers. Anthropic says all its bots follow robots.txt.
Bots keep a copy of your robots.txt for a while. OpenAI, Perplexity, Meta and Amazon mention about 24 hours; DuckDuckGo says up to 72 hours for DuckAssistBot. This checker also keeps a result for 15 minutes.
That is your choice. Blocking learners such as GPTBot or ClaudeBot opts you out of future training and does not hide you from AI search. Blocking searchers such as OAI-SearchBot or PerplexityBot does remove you from those answers.
Something in front of your site turned our test request away: a firewall, a security plugin or a CDN bot setting. Check those first. Keep in mind that our test can't come from the bot's own addresses.
Yes. We send about 16 small requests, at most 3 at the same time, only when you press the button. We keep the result for 15 minutes so a second check doesn't contact your site again, and we don't store your robots.txt after that.
Every bot in our list, checked against each company's own documentation. For ByteDance, Cohere, Diffbot and Omgili we found no company documentation, so we show their names as unverified.
The GEO check scores one page on 22 points that matter for AI answers, from structure to facts.
Run a GEO check